From aid to global development policy - United Nations · From aid to global development policy José Antonio Alonso Introduction Over the last decade, the international community

From aid to global development policy 11

Aid effectiveness a blurred literature

The most recent studies on aid effectiveness although carried out in a relatively similar theoretical framework vary in relation to the control variables used in the convergence equations the way of handling the endogenous nature of aid and the assumption of linearity of aid In terms of this last element the possible existence of decreasing returns on the aid was admitted in some cases while in others it was supposed that the impact of aid was conditioned by other characteristics in the recipient country

Among the studies in this last group the influential work of Burnside and Dollar (2000) stands out These authors believe that aid effectiveness depends crucially on the institutional framework and policies put into place by the recipient country After including an interactive term between aid and policies in the growth equation the authors confirm that the aid coefficient is not significant and that correlation to the interactive term is positive and significant Based on this study the World Bank (1998) concluded that it is important to focus aid resources only on those countries that enjoy a proper policy framework (although the real meaning of a universal ldquogood policy frameworkrdquo is debatable)

Other authors followed the approach of Burnside and Dollar (2000 and 2004) and incorporated an additional variable related to the recipient countryrsquos selected specific circumstances Among the factors considered authors include an export price shock in the affected countries (Collier and Dehn 2001) the degree of vulnerability of the economies (Guillaumont and Chauvet 2001 Chauvet and Guillaumont 2004) previous conditions of violence in the country (Collier and Hoeffler 2004) political instability (Chauvet and Guillaumont 2004) the level of democracy (Svensson 1999) or the limited size of the government (Economides et al 2008) In all these papers the effectiveness of aid conditioned by the recipient countryrsquos circumstances is confirmed although the results are highly sensitive to the methodologies used in the respec-tive estimations (Roodman 2007a and b Easterly et al 2004)

Additional evidence of non-linearity of aid effectiveness is the presence of diminishing returns In all cases where an appropriate variable was incorporated the estimation confirms the existence of decreasing marginal returns on aid (Hadjimichael et al 1995 Durbarry et al 1998 Hansen and Tarp 2001 Rajan and Subramanian 2005 and 2008) Although it is debatable where the threshold lies it seems clear that after a certain level the accumulation of aid can end up having negative effects on the recipient country Many of the studies incorporate the existence of decreasing marginal returns stating that the effect of aid is not necessarily conditioned by policies (Hansen and Tarp 2001 Durbarry et al 1998 Lensink and White 2001 Easterly et al 2004 Dalggard and Hansen 2001 Dalgaard et al 2004)

This new generation of studies also opened up some new issues worth considering For example Lensink and Morrissey (2000) and Arellano et al (2009) showed that instability of aid flows negatively affects aid effectiveness In these studies aid effectiveness is related not so much to the characteristics of the recipient but to the donorrsquos methods of operation The importance of donorsacute behavior is also underlined by Minoiu and Reddy (2009) in an estimation of the long-run relation between aid and economic growth

Some categories of aid (like humanitarian aid) have no relationship at all to the recipientrsquos growth while others (like health or education expenditure) only produce positive effects in the very long-term For that reason Clemens et al (2004) refine aid flows considered in their analysis leaving only those elements that have an effect on growth within a relatively short time period After that data disaggregation the results point to a

1 2 DESA Working Paper No 121

positive robust relation between aid and recipient growth in the short-term (in periods of less than four years) That result is also independent of the quality of applied policies and other control variables

Some authors admit that aid can be effective in some places and not in others For example Dalgaard et al (2004) incorporate the percentage of the total land mass located between the Tropics in a given country as one of the explanatory variables Their results confirm that aid is not effective in countries located in the tropical zone but it has a positive effect on growth in the rest of countries Unfortunately as Roodman (2007b) highlights these results are highly dependent on the behavior of a limited number of countries In the same vein Herzer and Morrissey (2009) try to capture the heterogeneous effect of aid in different countries After confirming that aid has on average a negative effect on GDP they show that in about one third of countries analyzed (included in their sample) aid influences growth positively Out of the 19 variables tested as potentially explanatory of the divergent impact of aid on countries only 3 (namely government size religious tensions and the rule of law and order) were considered statistically significant

Among the most skeptical mdashand influentialmdash studies are those by Rajan and Subramanian (2005 and 2008) Despite the sophisticated instrumentation strategy used the authors cannot find any robust relation-ship between aid and growth5 They argue that the absence of a relationship is due to two potentially negative aid effects first the impact that aid has on the political and institutional climate of the recipient country by deteriorating governance (Rajan and Subramanian 2009b) and second the effect that aid has on the recipi-ent countryrsquos competitiveness due to Dutch disease effects (Rajan and Subramanian 2009a) Nevertheless a negative coefficient of aid is found only in two out of eight main panel data regressions that is to say instances when the coefficient is non-significant are more numerous than otherwise But as Temple (2010 4448) points out ldquoAn insignificant coefficient should usually be seen as absence of evidence not evidence of absence at least until the economic implications of a confidence interval have been exploredrdquo

Other authors derived their conclusions from meta-studies based on the results generated by the specialized literature Doucouliagos and Paldam (2008) review more than one hundred studies and conclude the following (i) the effect of aid on growth is positive but not significant (ii) aid has a small negative effect on savings and an equally small but insignificant negative effect on investment and (iii) the effects associated with conditional estimates (when the effect of aid is conditioned by other variable) could not be replicated The authors admit the pessimistic tone of their conclusions

Arndt et al (2009) carry out an application of the Rubin Causal Model on the most recent publica-tions on the effect of aid Their results suggest that the aid-to-GDP-ratio elasticity of growth is somewhere between 010 and 023 for the longest time periods That implies that the effect on aid exceeds its supposed contribution to generating capital stock (estimated by Rajan and Subramanian 2005) and also positively influences the evolution of total productivity According to the value of the confidence intervals the authors cannot rule out the possibility that the aid elasticity falls into the negative zone but that it is positive wherever the results are the most solid

These limited results of the impact of aid on growth have led some researchers to look for the im-pact of aid on specific social indicators of direct interest For example Dreher et al (2008) or Mishra and Newhouse (2007) focus on the impact of aid on primary enrollment and infant mortality respectively Masud and Yontcheva (2005) center on infant mortality and illiteracy rates while Gomanee et al (2005) takes on infant mortality In these cases a positive effect of aid has been more frequently established

From aid to global development policy 1 3

Aid institutions and taxation

As mentioned above the decreasing return of aid on growth can be explained among others by the negative implications of these resources on the quality of institutions of recipient countries Aid can encourage rent-seeking activities alleviate pressures to reform domestic institutions and keep undemocratic governments in power (Moss et al 2008) Empirical studies have contributed to consolidating the negative image of aid (this conclusion is shared by several more-or-less related studies including Braumlutigam 2000 Knack 2004 and Braumlutigam and Knack 2004) In the same vein Djankov et al (2008) believe aid has an effect similar to that of the ldquoresource curserdquo negatively affecting growth through a deterioration of governance conditions in the recipient country Alesina and Weder (2002) and Knack (2000) find a correlation between aid inflows and corruption in recipient countries while Svensson (2000) finds a negative correlation between aid and democ-racy In this framework the exceptions are Tavares (2003) who obtains a positive correlation (although in this case institutional quality is only approached through corruption) and Coviello and Islam (2006) who cannot detect any effect of aid on institutional quality once country effects that are time invariant are controlled

However several of the studies mentioned above raise two important elements for criticism First they do not specify a model of institutional quality that is acceptably complete and their specifications raise suspicion of omitted variables Second the cited studies consider the relationship between aid and the quality of the institutions to be linear despite numerous reasons to see the relationship as one of diminishing returns

In a previous estimation Alonso and Garcimartin (2010) argued that institutional quality can be explained through four main variables the countryrsquos level of income per capita the average number of years of education among the population the level of social inequality (through the Gini index with a negative sign) and the level of tax resources Subsequently in Alonso and Garcimartin (2011a) they included aid (measured as a percentage of GDP averaged over five years)6 into this explanatory structure If only ODA is included (in constant dollars or PPP) the coefficient is not significant (although it has a positive sign) however when squared aid is also incorporated (to reflect diminishing returns) the two are significant with aid having a posi-tive sign and squared aid a negative one In other words aid has a positive effect on institutional quality but subject to diminishing marginal returns so that beyond a certain threshold the total impact becomes negative

One of the arguments often offered to justify the negative effect of aid on institutional quality is the negative impact that foreign flows have on incentives for setting up a solid tax system in the recipient country But having a tax system is crucial for consolidating effective and accountable institutions and for moving a country from aid dependency to self-sufficiency (Tilly 1992 Moore 2009 and Kaldor 1963)

Traditionally studies that analyzed the effect of aid on taxes led to notably pessimistic results received aid contributed to reducing the tax level of the recipient (Heller 1975 Cashel-Cordo and Craig 1990 and Khan and Hoshino 1992) However more recently studies have been less conclusive and the empirical evi-dence is ambiguous For example Braumlutigam and Knack (2004) find a negative correlation between aid and tax effort while other studies are unable of finding any relationship at all between the variables (Ouattara 2006) Morrissey et al 2007) Teera and Hudson 2004)

However papers taking a more comprehensive approach seem to detect a positive sign in the relation-ship between aid and taxation For example Gupta (2007) analyzes the determinants of taxes in 105 develop-ing countries controlling for structural factors Through several specifications and models of estimation he

1 4 DESA Working Paper No 121

confirms that aid has a positive although weak impact on tax resources Similarly Brun et al (2007) starting from a database elaborated by the authors7 estimate the tax ratio from structural variables In their study the effect of aid on taxation effort is positive and significant the institutional quality is not significant except in two cases and the interaction between both variables is positive in one case (the quality of bureaucracy) but not in the other two Finally Clist and Morrissey (2011) also find a positive relation between aid and tax effort particularly in the last decade

Alonso and Garcimartin (2011b) also have similar results They estimate a structural equation of tax effort in which they take into account both the usual variables (such as GDI per capita structure of output trade openness or rate of inflation) and one which is less frequently used income distribution Their results suggest that the aid coefficient is positive in all cases per capita income and the Gini index are also both significant and with the expected signs (positive in the former negative in the latter)

Some authors have stated that the impact of aid can be conditioned by the quality of the institutions in the recipient country (Azam et al 1999) Proceeding in a manner similar to that of Brun et al (2007) Alonso and Garcimartin (2011b) incorporated the product of institutional quality by aid into their previous estimation The results of the estimation reveal that both the aid coefficients and those of the interactive term (aid by institutional quality) are significant This means that aid positively influences tax resources but its effect is not linear and depends on the quality of a countryrsquos institutions in a context of weak institutions international aid could negatively affect the tax effort

The overview carried out in the last two sections is sufficient to confirm that the existing literature on the relationship between aid and growth is far from conclusive Despite that it is plausible point out that (i) the effect of aid seems easier to detect in the short-term than in the long-term (Clemens et al 2004) although a long-term positive effect might also exist (Minoiu and Reddy 2009) (ii) second aid seems more effective in contexts where it contributes to relaxing obstacles to a countryrsquos advancement such as vulnerability to external shocks or internal conflicts (Collier and Dehn 2001 Guillaumont and Chauvet 2001 or Collier and Hoeffler 2004) (iii) third the relationship between aid and growth seems to be subject to diminishing returns to the extent that beyond a certain threshold negative impacts can take place (Hadjimichael et al 1995 Durbarry et al 1998 Hansen and Tarp 2001 or Rajan and Subramanian 2008) (iv) fourth there may exist national factors specific to recipient countries that condition aid effectiveness but it is not clear what those factors may be v) instability of aid flows negatively affect aid effectiveness (Lensink and Morrissey 2000) and vi) finally aid can positively influence the quality of domestic institutions and there is nothing to suggest that it necessarily affect taxes negatively (impact may in fact even be positive) (Gupta 2007 Brun et al 2007) nonetheless due to decreasing returns beyond a certain threshold aid effects on both institutional quality and on taxation can become negative (Alonso and Garcimartin 2011b) Taking all the results into account and in spite of the pessimism shared by certain commentators there are reasons to believe that aid probably has a positive impact on growth although not necessarily in all cases In any case empirical evidence rarely supports the claim that aid in a general sense is actively harmful

Aid effectiveness donor response

The studies presented in the previous section provide a hazy picture on aid effectiveness Neither the use of insights from microeconomic theory nor the use of increasingly sophisticated econometric methods has elimi-nated the controversies underlying the results of the studies discussed above Moreover even in cases where

From aid to global development policy 1 5

the effect of aid is positive that effect seems to be relatively minor which justifies the efforts to improve aid effectiveness by the international community Although efforts seem to be well directed there are some aspects that require fine tuning while others are still lacking and need to be addressed

The Paris Agenda some critical observations

Widespread doubts about aid effectiveness have motivated donors to begin a revision of their aid management practices The origins of this process can be tracked back to the document entitled Shaping the 21st Century approved by the DAC in 1996 but the drive for reform received a renewed boost at the end of that decade due to the revision of the highly indebted poor countries (HIPC) initiative the debates around the Millennium Declaration and the outcomes of the Monterrey Conference on Financing for Development This reforming spirit was translated into a series of high-level meetings organized by the DAC One of such meetings taking place in Paris five central principles were defined in relation to aid effectiveness (i) recipient ownership of development interventions (ii) alignment of donor strategies and procedures with strategies and management systems of the recipient countries (iii) harmonization among donors (iv) management through results both for donors and recipients and (v) mutual accountability These principles were revised and expanded at the summits held in Accra in September 2008 and in Busan in December 2011

Without a doubt this is an agenda that requires donors to change their traditional means of opera-tion The OECD-DAC has monitored the implementation of the agreed reforms but the picture that emerges from its latest report is ambiguous progress has been made especially in the use of reliable public manage-ment systems and in the coordination of technical assistance but commitments are still far from being met (OECD 2011a)

In principle the process that started with the Paris Declaration must be judged positively The insis-tence on the ownership of aid intervention by recipient countries the desire to give greater predictability to aid allocations the insistence on more ambitious instrumental resources (such as programmed aid) the promotion of coordination between donors and the use of the recipientrsquos management channels are all recommendations that stem directly from studies on aid-effectiveness

However these agreements are also the product of the framework underlying their adoption Two traits stand out here (i) first the decisions were mainly led by the donors even if recipient countries were invited to the debate mdash there is a problem with the representativeness of the forum (DAC) where the process originated and evolved (ii) second perhaps again due to the nature of the forum where these recommendations were adopted their assumptions are based on an excessively technocratic (and rather naiumlve) vision of global effects of aid relationships (Whitfield et al 2008) Both aspects are limitations that must be overcome in the future

To be more precise the main reform efforts made by the donors seem to have been directed at cor-recting two important problems that penalize aid effectiveness the existence of perverse incentives stemming from asymmetry of information within the aid chain on the one hand and high transaction costs from the proliferation of donors and fragmented interventions on the other

The first problem derives from the fact that the ldquoaid chainrdquo rests on a system of principal-agent rela-tions that condition the incentive framework in which the players act the donor (who is the principal) decides about the allocation of resources but has limited capacity to control the management of those resources for

1 6 DESA Working Paper No 121

which the recipient (the agent) is responsible (Gibson et al 2005 Martens et al 2002)8 The fungibility of aid is a manifestation of this problem

Addressing the problem of perverse incentives is not simple In the past donors sought a solution by imposing conditionalities as a sort of implicit contract between donor and recipient However this route has proved rather toothless In a context of imperfect information it is not plausible to expect an optimal contract design that would exhaustively define in advance all conceivable possibilities states of nature As a consequence a problem of time-consistency in donorsrsquo behavior was common as well as the process had costs in terms of recipient ownership of reforms (Collier 2007) Furthermore it is also debatable whether the donor is in a better position than the recipient to define responses to the development problems that the recipient is facing

The only possibility for progress on correcting problems of asymmetric information would be through stricter alignment of the goals of both parties in the transaction and that is precisely the meaning of the principle of ownership In this case the recipient supposedly defines its priorities more clearly and the donor identifies those areas in which it wants to contribute Steps toward this goal include drawing up national development strategies as a foundation for defining the priorities the use of programmed aid instruments to support national policies the use of a recipientrsquos own fiscal system to channel aid (through budget support) and the formulation of medium-term budgetary frameworks that focus the aid commitments (Mid-Term Expenditure Frameworks or MTEFs)

However the ways in which these responses have been handled are open to some criticism Four observations are worth highlighting here

1 First there has been a certain misconception of ownership which has in practice been reduced to the capacity of a recipient government to define a comprehensive development plan with a clear set of indicators to be generated by a functional statistical system for monitoring progress It should be assumed that ownership is political and not a technocratic process and it is a gradual and complex process rather than a simple binary outcome

2 Second an excessively naiumlve perception of social dynamics has supported the national develop-ment strategies The demand for national consensus as condition for aid support is simply misguided In socially fragmented societies public action is the result of a confrontation between visions and interest and of an institutionally channeled political struggle In that context political consensus is more an exception than a rule

3 Third the alignment of the donor with the goals of the recipient has tended to be understood as a merely functional problem of adapting priorities as if there were substantial match between the interests of the two parties involved The rhetoric of the principle of partnership rests on that assumption But within that dialogue there might be also conflict and divergence of interests (Whitfield et al 2008) The agreement therefore will be the result of a strategic game in which both parties (donor and recipient countries) are implied

4 Lastly this type of approach has gone hand-in-hand with donors using highly intrusive for-mulas of negotiation and control which have given rise to demands for monitoring and evaluation systems of aid use (Whitfield 2009) This type of requirement was softened in the second generation of Poverty

From aid to global development policy 1 7

Reduction Strategies (PRSs) but there is still excessive meddling by the donors However as a recent investiga-tion revealed effectiveness can be achieved by much simpler means (World Bank BMZ and GTZ 2007)

The second major problem of aid effectiveness concerns the high transaction costs of aid This prob-lem has become more acute due to the proliferation of players in the aid system the overlapping of their priorities and the growing dispersion of their action (Acharya et al 2004 IDA 2007) Donorsrsquo responses to this problem include the establishment of a more active coordination process and measures to ensure fuller integration of their initiatives in the budgetary process of partner countries The principles of harmonization and alignment are steps toward those goals

Here too we find a process that should be considered positively but which presents some limitations that are worth highlighting First the lack of trust seems to underlie donorsrsquo behavior which is reflected in their resistance to the introduction of more effective approaches to support budgets in a coordinated way In fact if donors were genuinely committed to improving coordination they will further support multilateral channels a which in general does not seem to be the case As a consequence and beyond their rhetoric donors prefer to preserve their control on the channels of aid provision and management (such as projects and programs)

Second despite the fact that some conditions have been eliminated aid conditionality remains in effect in many cases In the case of many LICs dependent on aid conditionality reduces their policy space While it may be understandable that donors want to shape the policy framework where they would allocate aid resources imposing conditions for disbursement constitutes a poor guarantee of an adequate execution of those agreed policies Evidence indicates that this approach has been counterproductive conditionality is not effective when it substitutes local commitment to reform

Lastly limited progress has been made on improving complementarity and the division of labor among donors Some donors have advanced towards a more selective definition of their geographical priorities but that process has been carried out on the margin of identifying complementarities with other donors Labour division has improved within individual countries (national complementarity) but there is hardly any progress in improving complementarity of actions among countries

Perhaps it is still too early to judge donorsrsquo efforts in that direction However there are some obstacles that may constrain action by donors

1 First there are no sufficiently developed analytical tools to unequivocally to match recipient country needs and individual donorsrsquo expertise in specific sectors or areas In fact there is a wide overlap in donorsrsquo spheres of action they score low in sector concentration (or high dispersion) of aid and high in simi-larity among them (see table 2) These are not the best basis for identifying complementarities among donors

2 Second in terms of geographical priorities strategic economic and historical factors come into play in donorsrsquo aid allocation that are difficult to subordinate to an indisputable technical criterion Thus establishing a coherent framework for identifying priorities which would allow for reducing gaps and overlaps among donors is not an easy task and would only probably achieve partial success

1 8 DESA Working Paper No 121

Table 2 Concentration diversity and similarity of sectoral distribution of ODA (2006-2008)

DONORSConcentration by sector

(Herfindal Index)Diversity by sector (Shannon Index)

Similarity Index by sector(DAC)

Australia 00440 06874 5364

Austria 01618 05789 4639

Belgium 00305 07574 5741

Canada 00222 07717 5748

Denmark 00289 07335 5421

Finland 00267 07499 5354

France 00859 06286 5238

Germany 00436 07378 6267

Greece 01177 05446 3555

Ireland 00415 06845 5353

Italy 00281 07516 5554

Japan 00329 07337 5316

Luxemburg 00259 07310 4434

Netherland 00411 07305 6348

New Zealand 00510 06914 5889

Norway 00240 07734 6464

Portugal 00767 05794 3425

Spain 00185 07981 5871

Sweden 00271 07510 6060

Switzerland 00554 06727 5275

United Kingdom 00477 06666 5646

United States 00476 06952 6158

DAC Countries 00172 08196

Non-DAC Countries 00240 07644 6557

Source OECDDAC OECDStat online database Available from httpstatsoecdorgindexaspx

Note Herfindal Index Normalized

n

nS

HIN

ni

i

ji

11

1)(1

2

minus

minus=

sum=

=

Shanon Index )(

)1(1

NLn

SLnSSI

ni

i

ji

jisum

=

==

Similitude Index [ ]DACi

ji

ni

iSSSimI min

1sum

=

=

=

with i sector of aid j donor country N number of sectors and

sum=

=

= ni

i

ji

jij

i

ODA

ODAS

1

with i sector of aid j donor country N number of sectors and

Note Herfindal Index Normalized

n

nS

HIN

ni

i

ji

11

1)(1

2

minus

minus=

sum=

=

Shanon Index )(

)1(1

NLn

SLnSSI

ni

i

ji

jisum

=

==

Similitude Index [ ]DACi

ji

ni

iSSSimI min

1sum

=

=

=

with i sector of aid j donor country N number of sectors and

sum=

=

= ni

i

ji

jij

i

ODA

ODAS

1

From aid to global development policy 1 9

3 Lastly recipients sometimes prefer to maintain relationships with several donors in order to gain some negotiating room and to benefit from the existing differences among donors Diverse aid providers can bring complementarities and varied expertise For example the Republic of Korea benefitted from the differences in the respective aid approaches of the United States and Japan and Central America in the 1980s positively explored the differences in the aid policies of the United States and the European Union with respect to their respective approaches to the conflicts in the region

To sum up while some progress may have happened in improving complementarity among donors the improvement has been limited and incomplete This disappointing outcome demonstrates the limited contribution that enhanced complementarity among donors has had in lowering aid transaction costs and improving overall aid effectiveness

Aid dependency

The Paris Declaration should also be criticized for being silent about one of the chief problems affecting aid effectiveness the high dependency that aid generates in many developing countries

As studies have shown aid seems to present decreasing marginal returns There are various (potentially complementary) reasons for this outcome Relative large flows of ODA can generate (i) an adverse effect on the recipientrsquos competitiveness (the ldquoDutch diseaserdquo effects) (ii) undue pressure on the capacity of the recipient to absorb the resources efficiently (iii) complacency about the need to carry out fiscal reforms or to raise domestic savings (iv) diverting recipientrsquos limited human and technical resources (particularly those in the central regional and local administrations) and (v) a negative effect on the quality of institutions and their domestic accountability

Data reveal that dependency is high in a wide range of countries There are about 40 countries in the developing world where in the last 5 years the share of ODA to GDP is above 10 per cent Fiscal dependency is even higher as generally that coefficient is multiplied by a factor of between three and four to determine the share of aid in public budget In these cases the national institutions are being pushed to direct more attention toward their relationship with international donors rather than toward the demands or requirements of their citizens with costs in terms of their accountability towards the nationals of that country

Clearly the suspension of ODA flows is not an efficient response to this problem It would entail costs far higher than those it is trying to avoid For some recipients ODA is a source of financing much needed social expenditures and it is currently difficult to replace Aid dependency however makes it necessary to (i) be more cautious about plans to increase aid on the understanding that more is not always better (ii) establish plans to gradually downsize aid where feasible while seeking alternatives to sustain and improve achievements in the society concerned (iii) pay greater attention to existing routes for mobilizing domestic resources in developing countries This would involve strengthening tax systems as well as tackling tax evasion and capital flight and curbing illegal flows and (iv) finally dedicate more resources towards the provision of international public goods related with development goals (such as health diffusion of knowledge financial stability or fighting against climate change among others)9

2 0 DESA Working Paper No 121

Changes in the international system challenges for aid policy

As we emphasized at the start while the aid system has changed the world has also evolved but more rapidly and more profoundly It is a good idea therefore to look at such changes in the international environment and to identify the challenges they present for the aid architecture

Heterogeneity in the developing world

International aid was born in the 1950s confident in its (largely shared) diagnosis of the problems characterizing under-development Since then developing countries have experienced very disparate growth dynamics increasing the heterogeneity among them Thus a unique and shared diagnosis for the developing world is no longer valid

In fact during the last five decades developing countries followed very different trends One group mdash countries located in East and Southeast Asia mdash have managed to drive a successful growth process and have been converging towards income levels observed in the developed countries For another group mdash the Least Developed Countries mdash the gap has increased and they seem to be caught in a poverty trap (Guillaumont 2009) Finally a third large group of countries is situated between these two extremes This group has also increased its internal heterogeneity as a result of very different growth dynamics of its members

In sum throughout the last half-century a double divergence emerged (i) the distance between the poorest and the richest countries (LDCs and high-income countries (HICs)) grew and (ii) heterogeneity among developing countries also increased (Ocampo and Vos 2008)

Figures 8a and 8b illustrate the summary presented above In figure 8a presents the GDP per capita of the richest country (converted to PPP) as a ratio of the per capita GDP of the poorest country (also expressed in PPP) it also presents the average per capita income of the richest decile as a ratio of the average per capita income observed in the bottom decile global distribution of income for the period 1950-200810 In both cases the ratios are increasing confirming that the gap between the richest and the poorest worldwide has increased It is worth noting that the rather fast increase in the ratios after the 1990s with a slight reduction in the last years Figure 8b illustrates the evolution of heterogeneity among countries which is measured by the coefficient of variation of countriesrsquo per capita GDP (converted in PPP) also for the period 1950-2008 The heterogeneity in the world economy experiences an increase after the 1980s while heterogeneity among developing coun-tries also increases and very intensively after the 1990s

Another indication of this increasing heterogeneity is offered by the World Bank countries classifica-tion In the late 1990s LICs (some 63 countries) represented 59 per cent of the worldrsquos population and close to 20 per cent of the aggregated world GDP when measured in PPP (see table 3) However over the period 1999 to 2011 a great shift occurred within the groups 26 countries abandoned LIC status and became LMICs while 2 LICs joined the UMICs Meanwhile 28 LMICs became UMICs and finally 6 UMICs became HICs As a consequence in 2010 LICs (only 35 countries) represented just 11 per cent of the world population and a minuscule 13 per cent of the world GDP (PPP) On the other hand between 1998 and 2010 MICs (94 countries in 1998 and 110 in 2011) increased their share in world population from 25 per cent to 72 per cent and their contribution to world GDP (PPP) rose from 23 per cent to 44 percent Thus the majority of the population of the developing countries no longer live in LICs (as in the past) but in the broader and more heterogeneous group of MICs